The Generative AI Race Has a Dirty Secret
wired.comEnergy source > Energy consumption.
Energy consumption is not a bogeyman and we'll need to consume vastly more energy if we want to make people richer, healthier and happier.
We need to pivot to cleaner sources, so the focus should be on that.
Would be great if the journos could stop demonizing the wrong thing.
ML training also seems like it ought to be fairly dispatchable, right? Pause the training if there isn’t excess wind or solar. We’ll want to overbuild on renewable power capacity, so that if it not at peak, the lights can stay on. Having a use for that excess power seems like it ought to help economically incentivize the building.
That’s all well and good to say, but we currently have no system to do that.
What we ought to have is a carbon tax. Then if some activity such as bitcoin or ai or driving or flying uses carbon, you pay extra. Whereas if you manage to fuel your activities with nuclear/wind/solar you don’t.
There are loads of complexities to this but simply saying “consumption is necessary for riches” while ignoring that we have not managed to lower our carbon usage is not quite hitting the mark.
Carbon taxes aren’t necessary to give solar and wind an advantage. More than half of ~55GW new power generation capacity in the US this year is going to be solar.
> while ignoring that we have not managed to lower our carbon usage
We have in the US, both total and per capita. Carbon per person is down nearly 25% since 2000.
People seem to make up statistics in their heads to match what they think is going on.
World growth is the only metric that matters. USA offshored many industries which use CO2.
World carbon growth has not peaked as you can see here: https://ourworldindata.org/co2-emissions
CO2 is also most analogous to an overflowing bathtub. Lowering the rate would be great if we can get there, but what we really need is to get to zero new CO2 emissions.
We are a long way off from that so every new energy use worsens the issue until we get there.
if your bathtub is overflowing you should unplug the drain; atmospheric carbon capture is not cheap, but not infeasibly expensive either, and will surely be necessary
I agree, but you can only stretch an analogy so far. Unplugging a bathtub is free, whereas capturing carbon currently is more costly than the energy used to capture it.
taken literally you seem to be saying that there are costs other than energy involved in atmospheric carbon capture, which is too obvious to be worth saying; your capex and non-energy opex costs can be low but they obviously can't be zero or negative
do you mean more costly than the energy obtained by burning it, because that's a small edit that converts your statement into an assertion that's at least coherent
if so, no, the cost is about an order of magnitude smaller; though coherent, your assertion is still wrong
we do currently have a system to pivot to cleaner sources; it's the photovoltaic module industry, which has undercut the price of fossil-fuel energy by almost a factor of two in much of the world, though not in polar regions
Please see this chart of global co2 emissions. They have not stopped rising
We have a system for adding solar energy on top of carbon emissions but not for lowering carbon usage, yet.
they will continue rising until the late 02020s because ramping up pv production is a relatively slow process and because transport mostly can't be electrified without syngas or something similar
of course syngas will happen but it's not an overnight thing, especially if what gets adopted is hydrogen instead of something like fischer–tropsch-sourced diesel
As the original poster noted, energy is useful. Carbon won’t decline unless electric beats it for baseload, with batteries or somesuch.
Merely being cheaper won’t do it. As long as more expensive sources are useful, and as long as solar is supply constrained, no reason that solar growth would cut into carbon usage.
'unless electric beats it for baseload' seems like an error; perhaps you meant 'unless renewables beat it for baseload'
the cheaper/useful/constrained trichotomy seems like a similar error; supply constraints manifest as higher prices, and the usefulness of a fairly fungible commodity like energy is entirely relative to its price
pv with batteries is already cheaper than fossil-fuel sources in much of the world, and largely as a result, fossil-fuel consumption for electrical generation is already slowly dropping in much of the world; this has already resulted in a rash of bankruptcies of coal companies
overall fossil-fuel consumption is not dropping, but will be within a few years
A photovoltaic industry that mostly depends on fossil fuels, doesn't it?
no, it mostly depends on energy
the production processes for the key materials involved, like silicon, aluminum, and glass, can use any convenient energy source
the exception is the carbon dioxide emissions from the carbon electrodes in aluminum smelting pots, but this is a tiny fraction of the total
> no, it mostly depends on energy
Sure, and where does most of that energy come from today, and how do you propose we replace it? Because that's the whole point, isn't it?
> can use any convenient energy source
Sure. Right now it's mostly fossil though, and it's not like photovoltaic is even close to changing that. Of course if you assume a theoretical clean energy to build your photovoltaic, then that's clean (in terms of CO2), but if you have that clean source, why do you need photovoltaic in the first place?
I think that's a real problem that gets ignored too often: fossil fuels are limited, and we seem to be around peak production right now. So fossil fuel will go down, that's pretty sure. The question is whether we will be able to compensate, and it's really not a given.
most of this energy today comes from fossil fuels. i propose we replace it primarily with photovoltaic energy. i am not sure how this could have been unclear from my comments above. this transition is already happening and will be essentially complete within a decade
(supplemented where possible with wind, which is cheaper where it's abundant)
but it's irrelevant where the energy comes from today when we're talking about what pv production is dependent on. it's not dependent on fossil fuels. it's dependent on access to cheap energy, and it also provides that cheap energy; the energy payback time on pv production has been under two years for decades now
if your concern is that there will be no market for additional abundant cheap energy once current fossil-fuel generation has been displaced, you should probably stop worrying about that, because that has never been a problem so far in human history
oil extraction is indeed around its peak right now, but coal reserves would last another century or more at current extraction levels. the issue is that it would be bad to turn the planet into venus. currently the world is on a path to, as you say, compensate, but it's possible that a sufficiently large disaster could halt that process
cf. https://archive.is/KMsTT https://www.economist.com/finance-and-economics/2023/02/13/w...
> i am not sure how this could have been unclear from my comments above.
It was clear. I am just saying that photovoltaic is very far from replacing fossil fuels.
> this transition is already happening and will be essentially complete within a decade
I guess let's see that. Also photovoltaic energy depends on the weather, it's not exactly something we can control (unlike fossil fuels or nuclear energy).
> that has never been a problem so far in human history
Well I think we can say that energy consumption in human history is largely correlated with fossil fuels, can't we? But I did not say that there is no market. I said that it's not clear whether we can make it work _at all_ without fossil fuels and without reducing our consumption.
> the issue is that it would be bad to turn the planet into venus.
Yep, but the biodiversity problem is not even related to climate change: we are destroying life on Earth just because we have the energy to do it. The fact that, on top of that, this energy is creating big problems that still have to happen (climate change) is just adding to the problem.
We should just drastically reduce our consumption, both because cheap energy is not a given for the future, and because we should care about not turning the planet into Venus, or Mars.
if it was clear then why did you ask
are you just trolling
That convenient energy source needs to be stable, which most renewables are far from.
there are some advantages to it being stable, but even today, pv plus batteries is still cheaper than coal in much of the world
If you ignore the fact that your pv panels plus batteries relied on fossil fuel for extraction/manufacturing/transport, right?
Now try to do all that without fossil fuels. Is that still cheaper, today? Pretty sure it's not (pretty sure it's not possible to do that with only pv energy right now, the infrastructure is just not there).
no, they don't; you are harping on irrelevancies
>> we'll need to consume vastly more energy if we want to make people richer, healthier and happier
I have frequently heard this argument made - often as here axiomatically.
No argument that changing energy sources is crucial. But I would not dismiss a plan to also reduce overall energy consumption while maintaining or improving quality of life indicators.
> I have frequently heard this argument made - often as here axiomatically.
Okay, you're dismissing the argument without proposing any counter arguments.
J. Storrs Hall lays out the case for the relationship between energy usage & quality of life.
What are your specific rebuttals?
It is not self-evident to me that QoL improvements are necessarily driven by increased energy consumption. Not having read J. Storrs Hall, I do not have any specific rebuttals to that.
When I have heard such assertions before, I dug a little and found it was based on extrapolation. Sorry, if that isn't implicit in yours or J. Storrs Hall's thinking.
Ultimately, any system that seems to grow exponentially will eventually be found to have been on an S-Curve. We can argue the timing of reaching the plateau after the fact, but it is certain that no system can continue to grow exponentially indefinitely. Nor should it have to.
Global human population for instance should start to reverse centuries of growth this century. In the case of population flattening, perhaps we can agree this is to the good. I would argue much the same about energy consumption.
I just looked at J. Storrs Hall page on Wikipedia and saw Kurzweil in the References there.
Would it be fair to call J. Storrs Hall a Singularitarian ? If so, it figures. We can argue the merits of such arguments, but it is a reach to state them axiomatically.
Not the OP, but fossil fuels are not unlimited, until now we don't have anything close to them, and anyway more energy consumption tends to go with more loss of habitat (talking biodiversity).
I think there is a case to make about reducing energy consumption.
I'm 100% in the "let's make clean energy too cheap to meter" camp. Get rid of anti-dumping legislation and let China subsidise as much solar panel production as they can! Build wind farms everywhere! Long distance HV transmission between climate zone should be so common that people don't keep thinking it isn't in use now! (And yes, it would be great if we can get fusion working)
But we aren't there yet. Until we are we need to both increase efficiency and increase renewables.
Google in particular should be praised for their work in subsidizing renewable energy generation[1] by guaranteeing energy purchases from renewable projects[3]. This guaranteed income stream is often what gets them off the ground, and the higher carbon-free energy usage from Google's data centers than the energy markets they are in is great to see[1].
However, that shouldn't preclude journalists from keeping that energy usage in people's minds.
[1] https://cloud.google.com/blog/topics/sustainability/5-years-...
[2] https://www.gstatic.com/gumdrop/sustainability/2021-carbon-f...
[3] https://cloud.google.com/blog/topics/sustainability/clean-en...
This used to be true, but changed dramatically.
The relevant term is „energy intensity“, measuring how much energy is needed per unit of GDP. It has been improving for quite a while now, see, for example https://www.aceee.org/blog/2015/12/what-energy-intensity-and...
That doesn't change the fact that first and cleaner renewable energy source is consumption reduction.
> That doesn't change the fact that first and cleaner renewable energy source is consumption reduction.
1) That's not a fact
2) That sentence doesn't make any sense grammatically
3) Consumption reduction = lower quality of life & health for poor people (the rich have 0 interest in consumption reduction... do you see the davos crowd turning in their private jets?)
Having a house with better insulation reduced my mother energy costs from 6k to 2k, with 2022 being a year where the electricity was more expensive.
She does not have weirdly distributed heat islands in her house, and supported this winter way better than past years.
She reduced her consumption by 3, increasing her quality of life.
"That sentence doesn't make any sense grammatically"
I'm sorry, still trying to learn a foreing language, thank for your caring advice.
"Consumption reduction = lower quality of life "
That's false, see LED lights.
Efficiency has it's limits and many industrial processes are already pretty close.
Of course there are limits, but new processes can be invented, it's just history. Electric cars are much more efficient than ICE ones. First civilian jet engine designs could be efficient up to a point, then they changed designs. Laser uranium enrichment is so much more efficient that USA classified it [1]. Transformers training is more efficient. And so on...
[1] https://en.wikipedia.org/wiki/Separation_of_isotopes_by_lase...
Assuming "cleaner" means more sustainable for the environment (not only carbon footprint, but considering biodiversity), then I think it's pretty much a fact: we are living a mass extinction right now, and that's really correlated to our energy consumption.
If it's purely about carbon footprint ("let's finish killing the biodiversity and make living on Earth as close as possible to what it would be to live on Mars"), then it's still debatable whether or not we could keep increasing consumption with the end of fossil fuels (and that's coming).
Doesn't matter. Humans are not going to accept the the massively sustained decrease in living standards that it would require, which means it is not a politically acceptable solution, and therefore not a solution.
Whereas spending some money to get us all nuclear might be. If not, we will have to spend more to metigate the effects of global warming.
If we really wanted to solve this quickly and efficiently, we could kill all humans.
This is the abstention-only of climate management.
Until the sources are clean I will blame the consumers that do so without regard for the source.
I think people are always looking to dismiss emerging technologies. AI is especially scary to many people that read newspapers (and this site); it's going to listen to press releases and write newspaper articles, it's going to illustrate them, and it's going to write the FactoryFactoryInterfaceAdaptor that serves the content to the public. So pretty much everyone involved in this discussion has to fight it or lose their career. (I think that's pretty far away, but I can't blame people for being paranoid here.)
What happens in these cases is that people look for reasons why the emerging technology is objectively bad, and then throw out arguments and see which ones stick. Electricity usage implies climate change, and that's a popular fear to play on right now. "This thing that's going to kill your career is also why hurricanes are stronger now!" Of course, that is only partially true; maybe AI isn't the problem and driving 60 miles a day in stop-and-go-traffic to go to work is what we really need to crack down on. But, that sort of thing is 70 years old and people are comfortable with it, so there isn't much persuasion left to do; people know it's bad and do it anyway. Meanwhile, climate change was a huge problem before the phrases "Bitcoin" or "deep learning" were ever coined, so it's kind of weird to blame them.
This sort of complaining is also easy. Sure, with more electricity users coming online, we worry about the cost going up. The cost going up is good for killing energy intensive activities that you don't like; eventually it won't be profitable to mine Bitcoins or train AIs. But the cost going up also affects other industries that we DO like; poorer people heating their homes can't absorb the cost, or industries like aluminum smelting can't absorb the cost. So now we want to distribute resources with some system other than "the highest bidder wins", and you know what that means! Time to involve the government! But the government is terrible at reacting to things like this, and even though regulation is obviously the globally optimal path forward, it moves too slowly to really make a difference. By the time we charge data centers more for electricity, AI will have taken your jobs and the East Coast will be underwater.
(Meanwhile, many data centers are powered by cheap renewable hydroelectricity anyway and so aren't causing climate change. They can build their AI training data centers near a hydroelectric dam, but people aren't going to move their family so they can get cheaper electricity or spend less energy commuting to work or whatever. I'm not sure where this all leads, but it does feel like appointing a boogeyman to blame instead of actually fixing the problem. This goes back to my opening argument about seeing which argument sticks. If you don't look into things too deeply, you have a nice set of rules that leads to an obvious action. I don't like Bitcoin because the existing system works OK. Bitcoin uses a lot of electricity. Some energy use changes the climate. The changing climate kill us all. So we'd better kill Bitcoin. A simple path from "I don't like it" to "kill it before it kills us". Get out your pitchforks!)
I kinda hate these articles about the carbon cost of compute- it's utterly short sighted, gpt3 is infrastructure that creates huge efficiencies - the cost is a bargain. Not to mention this is like early days and we're driving down the costs everyday.
Sources for the "creates huge efficiencies" part?
Isn't it the case that training a ML model consumes vastly more power than using the trained model for inference?
If so, the problem may be somewhat overstated. Training is a dispatchable power load. It can be scheduled for periods of surplus power in regions where surplus power is periodically abundant. Training a ML model could literally follow the sun.
At Big Tech scale, the same model is used for inference so many times that most of the energy is probably from inference.
Note that if you have surplus of power, maybe you want to use it more wisely :).
Perhaps, you are right about utility, but I can think of few power loads as well suited to dispatch.
Bauxite has to be shipped to Iceland to take advantage of their abundant geothermal and hydro power for Aluminium smelting. Data moves with much less friction.
Has someone already made a cryptocurrency that requires that miners train an AI to demonstrate proof of work?
Could be interesting. One way to do it would be to have a training set and a hidden test set, and mining a coin means arbitrarily high performance on the test set. In order to keep the test set hidden from miners (otherwise they could just memorize it) they would need to upload their weights to be validated.
The main issue I see is that many of these models are massive. Dozens of gigabytes. You’ll quickly DDOS your validator nodes before getting anything useful.
I kid you not, I read a scheme such a thing a year ago.
But now I cant find the source... video encoding mining seems to have gained more traction :/.
sigh Web4 hype train, here we come....
We're not done with Web3